The present disclosure relates to an image processing device, an image processing method and a computer readable recording medium.
In recent years, pixels and read circuits to read signals from the pixels in imaging elements, such as complementary metal oxide semiconductors (CMOS), are diminishing in size. Such diminishing has a problem of decrease in sensitivity and increase of various noises. To deal with the decrease in sensitivity, a sharing pixel structure for sharing a single read circuit among pixels to regard the multiple pixels as a single sharing block and read signals is used to reduce the area of the read circuit and increase the aperture ratio of each pixel (ratio of light receivers), thereby increasing the sensitivity.
On the other hand, noises that occur in imaging elements include dark current shot noises due to dark currents and random noises resulting from thermal noises, etc., in a read circuit. In addition to the dark current shot noises and random noises, there are noises in which the pixel values keep representing abnormal values and blinking failure noises in which the pixel values vary at each shot.
There is thus a technology to detect defective pixels in which such noises occur. According to the technology, difference values each between the pixel value of a pixel of interest and the pixel value of a pixel around the pixel of interest and in the same color as that of the pixel of interest are calculated. The difference values and a threshold are used to determine whether the pixel of interest is a defective pixel. When it is provisionally determined that the pixel of interest is a defective pixel, difference values each between the pixel values of adjacent pixels each around the pixel of interest and in a color different from that of the pixel of interest are calculated. The larger the number of adjacent pixels each in a color different from that of the pixel of interest and having difference values exceeding the threshold is, the more a defective condition variable representing a possibility that the pixel of interest may be a defective pixel is reduced to detect defective pixels accurately (see Japanese Laid-open Patent Publication No. 2006-148230).